There are three aims of this project. The first aim is to demonstrate that autism can be distinghished from mental retardation in infants one year of age. This will be accomplished by observing 1st birthday home videotapes of 75 3-4 year old children later diagnosed with autism, and comparing them to 1st birthday party videotapes obtained from 50 3-4 year children with idiopathic mential retardation and from 75 children with typical development. The second aim is to determine whether variations in early course of development in autism (early vs late onset) are related to individual differences in later symptom expression and brain function. It is believed that approximately 25 percent of children with autism show a pattern of normal development followed by regression in the second or third year of life. Such patterns likely reflect different underlying brain mechanisms in autism and may index subtypes of the disorder. To investigate this possibility, children with early versus later onset of autistic symptoms, as reflected in observations from 1st and 2nd year videotapes, will be compared to determine whether they differ in terms of their patterns of symptom expression and neurocognitive profiles at 3-4 years of age. The third aim is to study more broadly the longitudinal development of children with autism from infancy through early childhood with the goal of identifying key factors related to outcome in autism. Children with autism and the comparison group of children with mental retardation will be followed longitudinally during the preschool period. Analyses will examine how three primary factors - age of symptom onset, early neurocognitive profile, and intensity of early intervention - are related to outcome by early elementary school age (6-7 years).